Pedestrian crash reconstruction using multi-body modeling - CiteSeerX

PEDESTRIAN CRASH RECONSTRUCTION USING MULTI-BODY MODELING WITH GEOMETRICALLY DETAILED, VALIDATED VEHICLE MODELS AND ADVANCED PEDESTRIAN INJURY CRITERIA Lex van Rooij, Kavi Bhalla, Mark Meissner, Johan Ivarsson, Jeff Crandall University of Virginia, USA Douglas Longhitano Honda R&D Americas, Inc., USA Yukou Takahashi, Yasuhiro Dokko, Yuji Kikuchi Honda R&D Co., Ltd., Japan Paper Number 468 ABSTRACT This paper develops a method for studying pedestrian to car impacts through detailed multi-body modeling of various pedestrian anthropometries and vehicle types. The pedestrian models constitute a multi-body representation of the global joint kinematics and inertia for five representative body sizes. Advanced injury criteria are defined for the pedestrian lower extremities, knee, thorax and head. The vehicle model of a small family car is defined by a facet element mesh for the front-end and windshield of the car. The contact stiffness is variable over the location on the vehicle mesh and has been validated against experimental results and FE simulations of the EEVC impactor tests. The underlying structures of the hood are defined as rigid ellipsoids. The developed model is applied to the reconstruction of two PCDS cases with a small family car. Injury risk data was collected from the simulation model and compared to the injury outcome for the pedestrians involved in these two cases. Results of this study show that the detailed model can distinguish the injury severity for various body parts at impact locations on the vehicle. INTRODUCTION Pedestrian fatalities in the US reached 4,882 in the year 2001, while 78,000 people were injured in pedestrian to vehicle crashes (NHTSA 2001). This represents 12.9% of the total traffic fatalities and 3.9% of all injuries in traffic in that year. While pedestrian fatalities in the US have decreased by 16% since 1991, due to better education, smoother designs for improved aerodynamics and safer infrastructure, the vulnerable road user problem has grown larger on a global scale. Mackay (2000) estimates the total amount of traffic fatalities worldwide at 950,000 in the year 2000, while the World Bank (2003) states that 65% of all traffic fatalities involve pedestrians. Therefore, the total number of yearly worldwide pedestrian traffic fatalities is as high as 615,000.

More than 50% of all pedestrian injuries are caused by an interaction with the front of the vehicle as Figure 1 shows (Otte 2001). The bumper area generally causes lower extremity injuries, the hood edge area causes hip and abdominal injuries, while contact with the hood and windshield accounts for injuries to thorax, head and neck. Impacts to these vehicle areas are the focus of this paper, while the remaining injuries caused by an impact to the road or to any other undefined object are not investigated. 6.0 %

23.7 % 27.1 % 20.6 % 42.3 %

65.7 % (secondary & road impact)

Figure 1. Injuries sustained to pedestrians per vehicle region (Otte 2001). The injury distribution per body region denoted in Figure 2 clearly shows that most severe and life threatening injuries (AIS 5-6) are sustained to the head, followed by thorax, abdomen and spine. Less severe injuries (AIS 2-4) are in 37% of the cases sustained to the lower extremities and pelvis, while the head accounts for 35% and the torso and upper extremities for the remaining 28%. This illustrates the importance of focusing on injuries to lower extremities and head, and to a lesser degree to injuries to the thorax, abdomen, spine and upper extremities.

van Rooij

1

% Moderate Injuries AIS 2-4 Head

mechanical behavior of anatomical structures in response to impact loading. The US government conducted the Pedestrian Injury Causation Study (PICS) in the late 1970s, where data was recorded from the accident scene, the case vehicle and the medical reports (Jarrett 1998). In the 1990s the Pedestrian Crash Data Study (PCDS) was initiated in response to a modernized vehicle fleet. The PCDS database contains detailed reports on reconstructions of the crash. Data is recorded and analyzed based on the accident scene, the status of the case vehicle, medical records, police reports and interviews with people involved in the crash and with possible witnesses (Chidester 2001). From an in-depth analysis of real-world pedestrian to vehicle crashes it is possible to obtain a better understanding of the mechanisms that cause injury.

% Serious Injuries AIS 5-6

35

80

8 7 Thorax 9

37

Upper Extremity

46

Spine

76

Abdomen

Lower Extremity (including pelvis)

Figure 2. Pedestrian injury distribution per body region (Chidester 2001). The pedestrian population extends from toddlers to the elderly population. Results published by NHSTA (NHTSA 2001) from the Fatality Analysis Reporting System (FARS) and from the National Automotive Sampling System General Estimates System (GES) demonstrate that pedestrians of all age groups are at risk. Figure 3 shows that people 25 years and older account for 78.7% of all pedestrian fatalities. It also shows that 39.7% of the injured pedestrian population are 20 years or younger. Therefore, we can conclude that middle aged and elderly people are more likely to die in a pedestrian crash, while younger people and children are more likely to be injured in a pedestrian impact. Unknown

Injured Killed

>74

The European Enhanced Vehicle-Safety Committee (EEVC) Working Group 10 recognized the need for regulations in the design of the front structure of passenger vehicles and developed a standardized test method to evaluate this (EEVC 1994). EEVC Working Group 17 evaluated the previously developed test methods and proposed improvements to the test method based on new data from accident investigations, biomechanical research and experimental tests (EEVC 1998). Experimental test devices have been developed that represent the head, the thigh and the full lower extremity as Figure 4 shows. Adult Headform Child Headform Upper Legform

65-74

Legform

55-64

Age

45-54 35-44 25-34 21-24 16-20

Figure 4. EEVC WG17 pedestrian impactor subsystems.

10-15 5-9 30 and angular acceleration (rad/s2) αp < 3000

AIS 3

Ommaya (1988)

Fshear (N)

845

Tolerance level

Mertz and Patrick (1971)

Ftension (N) Fcompression (N) Nij - Ftens-int (N) Nij – Fcomp-int (N) Nij - Mx-int (Nm)

3290 4000 6160 6160 216

Dummy IARVs (an Nij value of 1 corresponds Eppinger et al. (2000) to a 15% risk of AIS Lund (2000) 3+ neck injury)

Peak lateral acc. of T8 (g)

45.2

Peak lateral acc. of T12 (g)

31.6

TTI(d) (g)

85

(VC)max (m/s) (to whole thorax)

1.47

Compression (%) (to whole thorax)

38.4

Force (kN)

7.48

Force (kN)

7.98

Axial Force Fz (kN)

10

Anterior-Posterior Bending Moment Mx (Nm)

524

1000

HIC [time window]

15 ms

Max linear acceleration (g)

85

25% probability of AIS Viano et al. (1989) 4+ Dummy IARV

25% probability of AIS Viano et al. (1989) 4+ 25% probability of AIS 4+ 25% probability of fracture Injury threshold Fracture threshold

Lateral Bending Moment My (Nm) 524

Tibia

Knee Ankle

FMVSS 214 (2003)

Viano et al. (1989) Cavanaugh et al. (1990) FMVSS 208 (2003) Messerer in Nyquist (1986), Schreiber et al. (1997), Miltner and Kallieris (1989)

Lateral acceleration (g)

150

40% risk for AIS 2+ lower leg fracture

EEVC WG 17 (1998)

Axial Force Fz (kN)

10.4

Failure threshold

Messerer in Nyquist (1986)

Tibia Index - Fcr (kN)

18.3

Tibia Index - Mcr (Nm)

450

Onset of injury

Crandall et al. (1999)

Bending Angle (degrees)

14.6

Shear Displacement (mm)

16

Injury threshold

Kajzer et al. (1997)

Inversion/eversion angle (deg)

28

25% risk of injury

Funk et al. (2002)

van Rooij

19